Understanding trends and variation in paediatric fracture management in England

Sarah Lucas, Patrick Aldridge & Andrew Hood

18 December 2024

Rationale


GIRFT report found some trusts were manipulating more displaced forearm fractures in the emergency department rather than in theatre. If more trusts adopted this, more conservative, approach to fracture management a significant amount of theatre time could be saved1.

Many uncomplicated paediatric clavicle fractures can be managed without x-rays2, and this is also considered true for toe fractures.

British Society for Children’s Orthopaedic Surgery (BSCOS) guidance suggests that no referral/follow up is required for many fractures of the clavicle, elbow, wrist and toes where there is no or minimal displacement3.

So, there is interest in understanding what the potential reductions in activity could be if trusts were to implement more conservative management of paediatric fractures and potential for an new mitigator in the NHP model.





  1. Paediatric Trauma and Orthopaedic Surgery. GIRFT Programme National Specialty Report. April 2022. https://gettingitrightfirsttime.co.uk/girft-reports/
  2. Lirette MP et al. (2018) Can paediatric emergency clinicians identify and manage clavicle fractures without radiographs in the emergency department? A prospective study. BMJ Paediatr Open. 10;2(1):e000304.
  3. Modifiable Templates for Management of Common Fractures. https://www.bscos.org.uk/public/resources.

Methodology


The study population included those:

- who had attended an emergency department/urgent treatment centre in England between April 2019 and March 2024 AND

- were aged 16 and under AND

- had a SNOMED code for closed fractures of toe, clavicle, elbow, forearm or tibia/fibula recorded


Emergency care dataset (ECDS) linked with records in the Outpatient (OPA) and Admitted Patient Care Episode (APCE) datasets.


The coding of fractures is not sufficiently detailed and reliable to determine specific fracture types, and thus what would be the appropriate treatment at an individual patient level. However, we could calculate the proportion of children with each fracture type that:

  • had an X-ray in the emergency department
  • received a follow-up appointment
  • had a fracture manipulated in theatre (only includes closed manipulations and not re-manipulations)
  • had a fracture manipulated in the emergency department

Most common fracture types (2023/2024)


SNOMED description

Number

Percentage

Closed fracture of radius (disorder)

Forearm

48,043

35.2

Elbow fracture - closed (disorder)

Elbow

18,613

13.6

Closed fracture of radius AND ulna (disorder)

Forearm

17,498

12.8

Closed fracture of clavicle

Clavicle

15,259

11.2

Closed fracture of phalanx of foot (disorder)

Toe

15,232

11.1

Closed fracture of tibia (disorder)

Tibia/Fibula

8,549

6.3

Closed fracture of ulna (disorder)

Forearm

4,347

3.2

Closed supracondylar fracture of humerus (disorder)

Elbow

3,394

2.5

Closed fracture of fibula (disorder)

Tibia/Fibula

3,278

2.4

Closed fracture of tibia AND fibula (disorder)

Tibia/Fibula

2,145

1.6

Closed Monteggia's fracture (disorder)

Elbow

108

0.1

Closed Galeazzi fracture (disorder)

Forearm

65

0.0

Closed fracture of distal end of radius (disorder)

Forearm

23

0.0

Only showing those fracture types recorded 10 or more times


The majority of fractures are recorded under just a few snomed codes.


‘Closed fracture of radius (disorder)’ will likely include some proximal radius fractures that would be more accurately classified at elbow fractures.


Fractures of great toe were excluded, as these should be followed up. However these codes are not really used, suggesting great toe fractures may be coded as ‘Closed fracture of phalanx of foot (disorder)’ and thus included within our dataset.

Factors influencing whether a follow-up appointment is given

Odds Ratio

Confidence Intervals

P value

(Intercept)

2.20

2.15 to 2.25

<0.001*

Sex

Female

1.00

Reference

Male

1.12

1.11 to 1.13

<0.001*

Age

5-10 yrs

1.00

Reference

0-4 yrs

0.96

0.94 to 0.97

<0.001*

11-16 yrs

1.15

1.13 to 1.16

<0.001*

Ethnicity

White

1.00

Reference

Asian or Asian British

1.05

1.03 to 1.08

<0.001*

Black or Black British

1.14

1.1 to 1.19

<0.001*

Mixed

1.01

0.98 to 1.05

0.34

Other Ethnic Groups

1.00

0.97 to 1.04

0.86

Missing/Unknown

0.97

0.95 to 0.98

<0.001*

IMD Quintiles

1- Most deprived

1.00

Reference

2

1.08

1.07 to 1.1

<0.001*

3

1.03

1.01 to 1.04

<0.001*

4

1.05

1.04 to 1.07

<0.001*

5- Least deprived

1.08

1.06 to 1.1

<0.001*

Department type

Major Emergency Department

1.00

Reference

Urgent Treatment Centre/Walk in centre

0.96

0.95 to 0.98

<0.001*

Day of the week

Week

1.00

Reference

Weekend

1.06

1.05 to 1.08

<0.001*

Time of day

Day 7am-7pm

1.00

Reference

Night 7pm to 7am

1.12

1.1 to 1.13

<0.001*

Time of year

Autumn

1.00

Reference

Winter

0.94

0.93 to 0.96

<0.001*

Spring

0.95

0.94 to 0.97

<0.001*

Summer

0.99

0.97 to 1

0.06

Year

2019/20

1.00

Reference

2020/21

0.85

0.83 to 0.86

<0.001*

2021/22

0.79

0.77 to 0.8

<0.001*

2022/23

0.72

0.71 to 0.73

<0.001*

2023/24

0.72

0.71 to 0.73

<0.001*

Fracture type

Clavicle

0.79

0.78 to 0.8

<0.001*

Forearm

1.00

Reference

Elbow

2.47

2.43 to 2.52

<0.001*

Tibia/Fibula

2.09

2.05 to 2.14

<0.001*

Toe

0.37

0.36 to 0.37

<0.001*


Children are more likely to be given a follow-up appointment if they are

  • male

  • 11-16 yrs old

  • from an asian or black background

  • living in a less deprived area

They are also more likely to have a follow-up appointment if they attended

  • an emergency department

  • on a weekend

  • at nighttime

Those attending in more recent years were less likely to have a follow-up appointment, further indicating there has been a move towards fewer follow-up appointments.

Summary of management of fractures


Over the last 5 years

  • No change in the proportion of fractures being x-rayed.

  • Slight reduction in the proportion of fractures being followed up, but significant increase in the proportion that are conducted virtually.

  • The proportion of forearm fractures manipulated in theatre has decreased while the proportion manipulated in the emergency department has increased. The overall manipulation rate for forearm fractures has reduced over the last 5 years.

Overall, there is a trend towards more conservative management of paediatric fractures.

Opportunities to reduce hospital activity

X-rays for Clavicle and Toes fractures by trust (2022/23)


Percentage of clavicle fractures x-rayed

Min

5.9 %

1st quartile

87.7 %

Median

93.2 %

3rd quartile

95.7 %

Max

100 %

Reducing the percentage of x-rays to the level of the lowest decile of trusts (77.3%) would give an annual reduction in England of 1,986 (15.5%) x-rays.

Percentage of toe fractures x-rayed

Min

2.4 %

1st quartile

72.6 %

Median

82.4 %

3rd quartile

89.7 %

Max

100 %

Reducing the percentage of x-rays to the level of the lowest decile of trusts (62.3%) there would give an annual reduction in England of 2,412 (22.5%) x-rays.

Upper limb fractures with follow-up by trust (2022/23)

Min

19.8 %

1st quartile

56.2 %

Median

65.1 %

3rd quartile

77.2 %

Max

98 %

Reducing the percentage of follow-ups to the level of the lowest decile of trusts (46%) would give an annual reduction in England of 13,768 (30.7%) follow-up appointments.

Min

25.9 %

1st quartile

74.2 %

Median

83.6 %

3rd quartile

90.4 %

Max

100 %

Reducing the percentage of follow-ups to the level of the lowest decile of trusts (61.8%) would give an annual reduction in England of 4,486 (25.8%) follow-up appointments.

Min

5.6 %

1st quartile

46.5 %

Median

61.2 %

3rd quartile

82.2 %

Max

97.4 %

Reducing the percentage of follow-ups to the level of the lowest decile of trusts (31.4%) would give an annual reduction in England of 4,440 (49.9%) follow-up appointments.

Lower limb fractures with follow-up by trust (2022/23)

Min

20.2 %

1st quartile

72.8 %

Median

83.7 %

3rd quartile

90.1 %

Max

100 %

Reducing the percentage of follow-ups to the level of the lowest decile of trusts (60%) would give an annual reduction in England of 2,817 (26.4%) follow-up appointments.

Min

11.1 %

1st quartile

31.7 %

Median

45.4 %

3rd quartile

56.9 %

Max

100 %

Reduced the percentage of follow-ups to the level of the lowest decile of trusts (21.3%) would give an annual reduction in England of 3,260 (53%) follow-up appointments.




Total annual reduction of 28,771 follow-up appointments in England.

This may be an underestimation as number of children may have more than one follow-up appointment that could be deemed unnecessary.

Number of clavicle fracture follow-ups by trust (2022/2023)



This includes all outpatient attendances, including physiotherapy appointments, in the 3 months post-fracture.

There is significant variability between trusts in the number of follow-up appointments for clavicle fractures.

Many clavicle fractures should not require follow-up yet some trusts are averaging 2-3 follow-up appointments per clavicle fracture, indicating a potential to further decrease follow-up appointments if second and subsequent appointments are considered.

Summary of the potential activity reductions


Clavicle

Elbow

Forearm

Tibia/Fibula

Toe

Total

Reduction in x-rays in emergency care

1,986 (15.5%)

-

-

-

2,412 (22.5%)

4,398 (18.7%)

Reduction in follow-up appts

4,440 (49.9%)

4,486 (25.8%)

13,768 (30.7%)

2,817 (26.4%)

3,260 (53%)

28,771 (32.7%)

  • Scope to reduce x-rays for clavicle and toe fractures in England by almost a fifth.

  • Scope to reduce follow-up appointments for elbow, forearm, clavicle, tibia/fibula and toes fractures in England by a third.

Differences between our approach and the GIRFT metric


We have also considered manipulations but there are some significant differences between the data included by GIRFT/Model Hospital and the data used in this study. Our analysis:

  • excluded certain fracture types, e.g open fractures which are assumed to all require treatment in theatre.

  • excluded any activity coded as re-manipulations.

  • used only the most recent year where follow-up data is available (2022/23)

  • links emergency care data to inpatient data and so we are only considering manipulations in theatre for patients identified in the emergency care dataset as having a fracture.

GIRFT metric does not account for any regional differences in fractures rates, but it is unaffected by the poor recording of diagnoses codes in emergency care.

Our measure accounts for regional differences, but does rely on the assumption that the fractures requiring interventions and those that do not are equally likely have a diagnosis code recorded in emergency care.

Comparing methods for calculating the rate of manipulations

Graphs show our data for the number of forearm fractures manipulated in theatre in 2022/2023 (identified by linking to emergency care fracture records), but use different denominators to calculate the rate.

Top figure shows the 20 trusts with the highest (red) and 20 trusts with the lowest (green) rates of manipulations in theatre using total A&E attendances as the denominator.

Bottom figure uses the number of A&E attendances for forearm fractures as the denominator, and the same trusts labelled above are shown in their new positions according to this new metric.

Example 1, using total A&E attendances as the denominator The Royal Cornwall Hospitals Trust has the 11th highest rate of forearm manipulation in theatre in England, but the trust moves out of the worst performing quartile (to 43rd highest) when using the number of forearm fractures as the denominator.

Example 2, Conversely Chelsea and Westminster Hospital NHS Foundation trust doesn’t perform as well as first thought moving out of the lowest quartile (16th lowest) to the 42nd highest, with a similar rate to The Royal Cornwall Hospitals Trust.

However, overall the majority of trusts remain within the same quartile regardless of the way the measure is calculated.

Limitations

  • It is possible some of our numbers may be underestimates due to some issues with coding/reporting especially as diagnoses codes are not always recorded in the emergency care dataset. For example the number of manipulations of elbow fractures in theatre is considered low, this may be a coding issue or related to the inability from the coding to determine which radial fractures should be classed as elbow rather than forearm fractures.

  • Our incidence rates are for closed fractures only, open fractures, pathological fractures and birth trauma fractures were excluded.

  • Coding of fractures is not specific enough to determine at an individual level which fractures could be managed more conservatively, so we are relying on comparing proportion between trusts. Some fractures will be more complex and require manipulation and follow-up, but we can’t be sure whether all trusts have a similar proportion of more complex fractures.

  • Only closed manipulations without internal fixation are included in our data, so if some trusts are treating a higher proportion of fractures with internal fixation then their rate of closed manipulations could appear lower. We have also not included re-manipulations in our data.

  • There will be a small number of cases where a child has more than one fracture or other injuries.

  • It should be noted is that data is allocated to the trust where the child attended the emergency department, but where they were followed-up if that is different